Homemade 15 emitter high-CRI Rebel LED flood light with red and cyan boost

[jason 77]

Jason 77, your schematic looks fine EXCEPT - (1) I would get rid of the .001 capacitor at the sense resistor, and (2) check the voltage at the positive input of the comparator with the potentiometer set to max. .05 volts gives you 1.0 amp. I get .045 volts there at 850ma, for example. Also, I sent you a private message, not sure you check your forum messages... Jeff

Ok I rebuilt the circuit using a fresh LM393 and took out the .001uF as you suggested. I actually went down and bought a 10kohm linear pot from frys as I wasn't 100% sure the ones I have are linear. The circuit now works fine I think, I can touch the pot and the heat sink for the mosfet and there is no change in the light output. I took some measurements and found that when the voltage at the positive input of the comparator is .05 volts the circuit is suppling the LED with 800mA of power, however at full tilt "pot up all the way" I get .074 volts and the LED is getting 1.24 Amps... The .05 ohm sense resistor is rated at 1% tolerance but when I measure it with my multimeter "fluke" I get .03 ohms, although I am not sure how accurate the meter is when measuring resistance under 1 ohm....

The high-CRI light remains unmodified. But I have made a new light that helps to fill-in with my photo lighting setup, Hydra-3. It is a non-high CRI neutral white Rebel-ES 1000-lumen flashlight and 2,000 lumen floodlight. It does have a light-stand fitting like the high-CRI light:
Jeff O.

I am curious as to which Rebel ES LEDs you are using as the highest neutrals I have seen are rated at 310 lumens at 1 amp?

Jason 77, if you haven't given up, allow me a silly question. You do realize that you have to supply ground and V+ to the comparator, right? These connections are not shown in the schematics reproduced in this thread.

Yes I do realize that thanks for the hint though...

 

[jason 77]

testing as my last reply isn't showing up.....

 

[anuragwap]

deleted

 

[jeffosborne]

Hi anuragwap! The output of the LM239 comparator is open collector, which means it will only sink a circuit to ground when the output is low. It will not pull the circuit high when the output is high, so a pull up resistor is required. The 100K value is a very mild pull up, as the input of the mosfet requires little to no current, only a voltage. A lower value resistor would work, for example, but it would needlessly drain current when pulled low.

Your 8 18650's would be 33.6 volts fully charged, so be sure to check the specs on the mosfet transistor you use. The maximum gate to source voltage for the mosfet I used is 20 volts as you mentioned, so a resistor and zeiner diode could be used to limit the gate voltage. Also, because this is a linear drop down regulator, it is most efficient to keep the supply voltage just above the total voltage required by the string of LED's you are driving in series. This makes for less heat at the mosfets, too.

I did find a hot new mosfet that has a super-low on resitance: 1.2 milliohm at 195 amps! It costs a little more, but super performance. It is an International Rectifier IRLB3034PbF. I think they were $1.67 each at Future Electronics.

So what do you have in mind for your light?
Cheers,
Jeff

 

[anuragwap]

deleted

 

[Waldur]

Hi!

Sorry for bringing up the old topic, but it is too good for being forgotten
As I am building a PSU for a bunch of LEDs (driven by 700mA each) and I am allowed to use only 5VDC voltage-line, a low-voltage-drop driver is more than perfect!
I'm not into inductors, and chips that should be soldered in the oven, I find this circuit with comparator very attractive.
If you can lead me to the final solution, I will be more than pleased
This currently used FQP50N06L is very good for me, it is available and with good price.
I will power the driver up with 5VDC from the regular power supply. And use only one Cree XRE 3W LED. Voltage drop across the LED is approx. 3.5V.
There is no need to dim the brightness manually, it means no pot will be necessary. I will test the circuit and get the right values for the right current in the output.
But what do I need is to get the possibility to turn the circuit/LED on and off via uC output, and by that I mean the controlling voltage is 5VDC, but with very low current.
One way is to use ULN buffer, in parallel, and turn the whole circuit on and off.
Another way is to keep the circuit running and let the signal from uC to the right place in the circuit to set it on or off. And maybe not use additional components at all? OK, the uC will be Arduino UNO, but using just digital I/O's to switch on and off.
It would be like a PWM signal-input, but in my solution used only for 0% or 100%.
Is it possible to achieve that?
I believe that someone has this solution in his/her head, and if it possible to get it finally drawn down, the X-mas will be much whiter

Thanks in the advance!!!!



All the best,


Heldur